Skip to content

[dev] docs: v1 roadmap draft#32

Merged
sherodtaylor merged 7 commits into
mainfrom
feat/roadmap-v1
May 26, 2026
Merged

[dev] docs: v1 roadmap draft#32
sherodtaylor merged 7 commits into
mainfrom
feat/roadmap-v1

Conversation

@sherodtaylor
Copy link
Copy Markdown
Owner

What was requested

Sherod asked for a v1 roadmap from his candidate list with critical feedback.

What changed

Adds docs/roadmap-v1.md:

  • Ranks each candidate by leverage-per-effort (ship now / ship soon / defer).
  • Calls out three gaps the original list missed: per-agent capability scopes, cost/budget controls, proactive work origination.
  • Sequencing recommendation (v1.0 → v1.1 → v1.2 → v2.x).
  • Four open questions at the end for Sherod to weigh in on.

How to verify

Read the doc; this is a draft, not a commitment. Iterate via PR comments.

🤖 Generated with Claude Code

devbot and others added 7 commits May 26, 2026 02:02
…s list

Critical pass on the proposed v1 candidate list: ranks each item by
leverage-per-effort, names the items that don't earn their weight yet, and
adds three gaps the original list missed (capability scopes, cost budgets,
proactive work origination).

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Restructures the doc so the reasoning flows from purpose → gap analysis →
themes → features, instead of starting in the middle with a ranked list of
Sherod's candidates. Every feature now traces back to one of five named
promises (P1-P5), and the explicit non-promises section reduces drift risk.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Homelab framing lowers the bar — implies hobby, throwaway, "good enough
for me." Production sandbox names the right stakes: real cluster, real
infra, real PRs, real blast radius. What makes it a sandbox is the
boundary (single operator, isolated network, recoverable state), not the
quality of the work. v1 has to clear the production bar, not the hobby
bar.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
… for it

Removed the "homelab vs production sandbox" editorial from the vision
section. The framing belongs in public-facing positioning (README, repo
description), not as in-line commentary inside the roadmap. The doc now
just speaks confidently from that posture.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- Tagline: "for a k8." (truncated) → "An autonomous engineering crew that
  ships real work against a real Kubernetes cluster."
- "Problem this solves" rewritten to frame the work as real engineering,
  not routine homelab chores. Surfaces the production-grade substrate
  (GitOps via Flux, Infisical secrets, VictoriaMetrics/Logs, egress
  capability scope) so readers understand bots ship to main, not toys.
- Replaced "for a homelab answering dozens of messages" with "for an
  always-on crew" in the Claude Code CLI rationale.

Technical references to the homelab repo, namespace, and manifest paths
are kept — those are accurate names, not framing.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
…ggers

Replaces "What v1 explicitly does NOT promise" with "Future
considerations" — same content, forward-looking framing. Each item now
names the trigger that would pull it into a future version (model
becomes load-bearing for cost/latency; crew grows past ~5 agents or
spans trust boundaries; bots take untrusted input or app-layer logs
prove insufficient). Appendix verdicts and the v2.x sequencing line
updated to match.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
…re consideration

Names channel-pluralism explicitly: we're building a framework, not a
Matrix-only tool. The agent loop is channel-agnostic — CLAUDE.md persona,
MCP tooling, dispatch, capability scopes, and run-id correlation all
apply regardless of where the human message enters. Trigger to pull
forward: a non-Matrix user need, or a second channel plugin landing in
the Claude Code marketplace. Slack flagged as the likely first add.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant